VoxeINET's Geo-Located Spatio Temporal Softbots: including living, quiet and invisible data

摘要

Linnaeus and Darwin understood the need to classify 'living things' to determine the basis of their relationships and interrelationships. In an Internet-of-Things (IoT) world we need to do the same, to be able to identify and compute with objects by type. However, the IoT does not inherently deal with spatial or geometrical structure, and mass phenomena (e.g. air, water, rock) are not objects per se. This can exclude these 'non-object' things from the IoT, which can be a severe disadvantage in many application domains. The solution to this is voxelisation of mass phenomena in the world within an overall coherent three dimensional coordinate reference system. This allows 'non-things' to be coherently situated, classified and treated computationally in the same way as discrete things and individual objects. VoxelNET is a distributed digital architecture that supports this voxelisation model, providing a world of voxels containing various information at different geo locations that can be compared in terms of numerous and unlimited taxonomical categories, and over time. Performing computations across this highly distributed system of systems can greatly benefit from the use of distributed softbots or agents without the need for centralized computations or control. Hence the VoxeINET distributed architecture not only parses objects and materials into computable objects, but also includes spatially located and volumetric computational agents that can collectively achieve analytical outcomes in an inherently distributed way. Here this approach is exemplified by distributed VoxeINET agents collaborating to conduct 3D volumetric path finding through the VoxeINET space, using a distributed Dijkstra pathfinding algorithm. Stronger implementations of the agent concept can include supplementing the basic Dijkstra algorithm with more sophisticated competitive and/or collaborative behaviours on the agents/voxels involved.